PURPOSE Gliomas are aggressive CNS tumors with significant heterogeneity, posing challenges for effective treatment. This study aims to enhance glioma classification by integrating multi-omics data, including genomics and magnetic resonance imaging (MRI)–based radiomics, focusing on metabolic and immune subtypes. METHODS Transcriptome data from 1,720 patients with glioma were analyzed to identify key prognostic factors, including 42 metabolism-related genes and 25 immune cells. A metabolism-immune classifier was developed to categorize gliomas into four subgroups: Metabolism high /tumor microenvironment (TME) high , Metabolism low /TME high , Metabolism high /TME low , and Metabolism low /TME low . Multicohort MRI radiomics combined with machine learning algorithms were used to predict these subtypes. Single-cell RNA and spatial transcriptome sequencing were used to validate subgroups' metabolic and immunological characterization. RESULTS The Metabolism low /TME low subgroup showed the best prognosis, whereas the Metabolism high /TME high subgroup had the worst. Machine learning models can predict glioma subtypes noninvasively based on MRI radiomics. Single-cell RNA sequencing confirmed the distinct metabolic and immune profiles of the glioma subgroups, revealing significant cellular heterogeneity within the TME. CONCLUSION This study demonstrates that integrating multi-omics data with MRI radiomics provides a robust framework for glioma classification, enabling more precise and personalized treatment strategies. The findings highlight the critical role of metabolic and immune profiling in understanding glioma heterogeneity and improving clinical outcomes.
Li et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: